Skip to main content

Full text of "Color removal from a neutral sulfite waste using magnesium coagulation"

See other formats





I^UTaan aM J. Edward Sinfley whosa technical and eaaniplarx contri* 

sincerely appreciated. 1 wish Co extend my appreciacion to H.F, 
Berfer, who, through the Haclonel Council for Air and Screojn lvprove> 

his cooperation and patience in high osteeB. The eoncribuced research 
and extensive laboratory work by Gary Christopher and Eevin Beaudec in 

I an sincerely grateful for the sacrifices ray wife Js 
bear during ray research. 

I have had aany rewarding experiences at Che Uaiversicy of 
Florida end am graceful for the opportunity co have been pare of 
that institution. 



1-1 Conersl Bsckground 
l‘Z LeE&I BequtreaontB 
1-5 Purpose ofThis Work 

ChareoterisUcs of Color 

ui Pulp Mill Effluents 

il Equipaent and Techniques 

>r HeasurcBsnt 

Coagulant Recycle 

Igss-— ■ 











Solubility diagraa ol 

) ugnesiuin rocovory s: 

$0.27/1000 to]. 



PuJp afld paper ■nnufacturing is one of the largest Industries 

in the United States. It is also one of thenaajor water using Indus- 
tries in the nation, produeing free sons allls extreealy large vol- 
umes of highly colored effluents, which are typically discharged to 
waterways. Color oreatos a unique problem in a stream. It is readily 
Idantifisble in an aesthetic sense end can detract fr^ the natural 
beauty of a body of water. The amount of light penetrating a stream 
would bo affected by a colored waste discharged to that stream, and 

!T companlas, haa recognized 


A highly simplified N5SC pulping process diagram is lllustratod 
mixed with B sulfite cooking liquor. The function of the sulfite 

RS. I.l 

cooking liquor is to sepBracn the lignin fron Che uoi 

pert of Che uaste effluent. Popcnding on the ulcinate use of Che 
pulp, eddltionel color reaovsl sief occur. Bleaching will further 
lighten the pulp. After each bleaching operetlon the pulp is washec 

using a NSSC production process. In the Fedonl Register under Pu: 

Standards, this law states in sunaary: All !^C plonts oaiat resovc 
7SS of their effluent color by 19S3, and ell new NSSC paints built 

1-3 Purpose of This Wark 

process for NSSC wsste snd to give Insight into the mechanlsa by 
which thet color was removed. The first objective was to develop 
a nethod of NSSC color removal that could be evaluated for use as 
a full-scale treatment process. Tho investigation was United to 

coagulant recovery on a laboratary tcale. The second objective was 
to investigate the mechanism by which color removal occurred. 

Techniques ea^loyed in 
relntionships developed 


vBvel«ngth of light otrihiag tho eye wll 

:hln the visible region of the 
jr perceived related to the wave- 

Obieccs ere seen by either cransnittod or reflected light. When 

passes through a mediim such as a solution of H5SC vastOi the nedlua 
appears colored to the observer. Since only the transaitted waves 

wavelengths of the spectmin depending on the olectroalc structure of 
the coDpound. A change could occur in the electronic configuration 

Very little evidence has been gathered on the naount of environ- 
aental degradation caused by color. Properties of pollutants such 


discharge of high!/ 

mild detinlCel)' effect the aesthetic 
Color would hove o derriaootal effect 

constituting shout or 

:h of the woodir tissue in plants. It 

natural fomation of this cross-linhed polyoeric saterial froa coni- 
ferpl alcohol and related substances is not presently compiecely 
understood. Despite considerable research, the structural characteri- 
zation of lignin has heen only partially successful. 

Freudenberg gethered infonDation about lignin structure 

from direct oxidation of lignin, froa bio-chaalcal experiacnts relatod 
to alcohols, and frca lignin degradation with strcpng alkali, methyls- 
tion and oxidstion. His exporiaents enabled an estiastion of the 
relative amount of alcohols which served as building blocks of lignin. 
Llgnlfication occurs in plant cells when alcohols are liberated and 

free TOdioals produced then coabine and build up lignin. Freudanberg 
(1966) formed a qulnoneacthide, as shown in Figure 2. 

in lignin. Since Che qulnnnraathldc has no opportunity tc 

for spnce lignin, uhich probably is siniUr to other wood lignins. 

revotl different ways in which Che units are combined. 

Through natural and industrial processes the lignin is separo' 

wood. The lignin was fractionated by molecular gels into three 
spectruB. Alder ec ai. (1966) degraded apruco lignin by acid roflun- 

ny investigators hs 
(1917) concluded through electrophoretic studies that nost 
nge. Block nod Christman (196ga) found that color collected 

Ftj. 22 


in different wntor saaples hnd eioildr chenienl and physical 

They devonstracod by dialysis that most of Che color 

Infrared spectmn for each of the fulvic fractions, the equivalent 
waishts of chose fractions, and the concentrations of the fulvie 
and huaic fractions in eaoh colored saaiple eore slBllar, Black and 
Chrlstaan (1963b) deoonstraced chat color intonslty uas pH dependent 
and would Increase with Increasing pK. They also found by dialysis 
that color existed ss a colloid, because only 10 % of the original 

Shapiro (iOSB) found chat organic color was aalnly dlcarboxyllc 

Shapiro (1998) demonstrated, by chromatographic ccanparisons, that 
pies were due to inorganic constituents of the water. Black (1960) 
the colloidal site range. 

Christman and Ohassemi (1966) isolated seven dlfferooc phenolic 

■olflcul&s Nitb hydm^l, 
Christun and Ghassomt ( 
soil would incrosse with 

ack snd Cbrlstasn<1966) , t 
Howaver, this increase 

Tsylor sad Zoltek (1974), i 
color r^ovol by Bsssive Ca(Ol) i 

light. Tho amount of color incraase in the soil-contacted saaples 
was directly proportional to the organic content of the soil- 
Glesslng and Samdol (19SB) studied color fluctuation in a chain of 
four Norvegien laies and found that color decreased in all of them 

had a high organic matter cootent. Gjossing and Samdal {1966} recor- 
ded a direct increase in the color of the impounded lake with time 
of water storage. Their data led to the conclusion that solubilired 
organic natter produced the color incraase in the impounded lake, 
and the degree of color Increase depended on time of impoundment. 

Peckham (1S64) separslod enlor from seven different waters into 
the ssme classes as did Black and Christman (1966). He found, based 
on filtration of the fractions, that tho ftilvic acid fraction Misted 

lar site range. Packhnm (1968) also revealed that both the fulvlc 

SOBt iJi I natursl »ot«r jel filtration and Pound molecular site 
distributions ranging froa greater than 200,000 to as low ns 700. 
Ihey found chat the laolecular site fraction that contained the lar- 
gest concentration of organic carbon did not produce the greatest 

liidwood and Felbech ClbOOl purified a yellow color fron organic 
Qaick and found that the organic Tsetter producing color was resistant 

The infrared spectra showed that aroaiatic carboxylic acids with ali- 
phatic side groups oontalning phenolie hydroxyl groups were eiejor 
eoBiponents of the color molecules. Day end Pelbech C1D74) obtained 
e yellow water-soluble organic exudate from the deoestic waste water 

noy help oliofnate color problems in watersheds. 

n [19711 end O'Melia (19721. In thie section th 

le iBfflediately adjacent to the colloidal parti- 

liquid Interface. These charjed ions are attracted to the colloidal 
surface electrostatically and repelled due to diffusion. The Verueey- 
Chrerbech aodel as described by Osipou fI972) stated that the London- 

te separate colloidal dcublo layers. Osipoa (1972) 

Quoy. Chapoan, Stem and Helaholta. The essence of the final Bodel 
Has that colloidal suspension uouid be destabilized if Che electrical 

concept vas supported in sc«e aysteais by the SchuUe-hardy rule, 

Hhiah acaces that the critical coagulation concentration of mono-, 
di- end trivalont ions to coagulate sols of the opposite charge are 
in the ratio of 100:1.6:0.13. Hatijevic ot al. (196da) developed 
a stabiliaaclon-dostabilization nodol for hgBr and Agl suspensions 

iona gained from the hydrolysis of AICNO^J^. Uotijevic « al. 
(196db) attributed the destabiUaation of the sols to the Al** spe- 
cies on the basis of charge reversal in the coagulation reaction. 

Houever, there was a stabilisation of the sol which waa followed 
by another sol coagulation. Hatijevic et al . (196Sb) contributed 

;h9 Tostabllized sol prior to Al(OH)j precipita- 

laMor (1607] dovoloped a bridging theoi 

daatabilitation of 

Mould serve as a bridge 
other oolloid vas svallable for 

and rostabilize the suspension; 4) 

Mould attach itself to the colloid 

suspension; 5) the aoount of eolloidel surface area present vas 
diroctl/ proportional to the aioount of polyaer required for coagula- 

betveen an anionic polrner and negative colloid would produce 

Packhaa (196B) studied coagulotioa of eight different clays by 
aiuminue hydrolysis and found the coagulant dose continually decreased 
with increasing concentration. Solubilized calclun and oagneslun 

biliae clay suspoosions. Packhaa (1966) deaonstratod that Che hydro- 
lysis products of Blun were important to clay destabilization by 

floe h&d the some teta potential for 

Apperantlf electrostatic forces were not controlling 
t [1968] studieO the CefloccuUtlon of eater 
sorping clays hy anionic and nonionic surfactants. Ho found that nati- 
Bun deflocculation was produced when the surfaces of tho clay lattices 
were cosipletoly covered with the nonionic surfactants. 

The American Chetustry Society [1968] puhlished data for aluBinum 
hydrolysis in colloidal suspensions showing that tho polynuclear spe- 
cies of slLBiimjn wore iaportont destobilitatlon factors. A colloidal 

the coagulant. They suggested Che forces of adsorption hotveen the 
colloids and the hydrolysis produets were responsible for dostablli- 

likely to aceuBulate at the solid-liquid interface. Another factor 
leading to colloidal destabilitation was that the hydrolysis products 
had more than one OH" ion that could sorp at tho interface. Their 
dace indicated that as Che colloidal surface art 

Langelier and ludwig (1949} experimented vl 
floeculacioo of four different turbid waters varying in exchange capa- 
city. They concluded that Che aechanisa of colloidal desieblliaation 
was controllod by the exchange capacity of the colloids. Michaels 

coagulation, found chat B aaall amount of hydrolysis was best suited 

for destabUimlao. He suggested that the destabiUzetion ■echanisB 
was a two step prooesst 1 ) polder sarptioa onto the colloidal sur- 
face and 2) interpareicle bridging following poZyner sorption to 
destabilize the colloidal suspension. 

Black et ai. ( 1965 ) evaluated coagulation by anionic polyners 
and demonstrated destabilization followed by restabilization with 
exoess polyner concentretion. Since both Che polymer and the colloid 
woro negatively charged, the destabilizstion was not attributed to 
coulombic forces, but to the build-up of Interpanlcle bridges 
through other than electrostatic mechanisms. They also found that a 
higher velocity gradient for a shorter time period was more effective 
in deatabiliza'ioa Chen a lower velocity gradient for a longer time 
period. Regunathen et al. (1975) treated turbid waters with alum 
and concluded, from zeta potential measurements, that the hydrolysis 
products of alum were controlling destabilization by sorption mechan- 

; et al, (196Ba) eraninad metal sorption onto a HnO. 
and found noutrel and anionic species did not sorp, a fact 

t al. (196Bb) studied Ci** sorption onto a negative MnO- sol and 

Ing value. The limiting Ca** sorption indicated a Langmuir aonolayor 
is probably oecurrlng on the WiOj surface. They restabllized the sus- 
snsion with more polymer addition, but did not achieve restabillzation 

nge for coagulation. 

m®)- 5u«est«i that a chsice of coa^aot aid would be bated oa the 
potential detemininj ions of tho sol, Robinson « el. (I974) rep„_ 
tod that Inrjet iooreases in tho turbidity of a river water inoioasod 
treatoont costs and large quantities of alias were required to produce 
potable water. Nonionic and cationic poiyelectrolytes were found to 
be me effective than aluu, suggesting that for this water an electro- 
static aechanisB was not controlling destabliiiation. Alianimm hydro- 
lysis was probably reMvlng turbidity by cnBealmont in a sweep floc. 

UMer C1967) defined coaguletlon as a kinetic process going Free; 
a quasi-stable to a note stable phase, and floccalation as tho 
bridging of already coagulated particles that entered into hindered 
settling. As on eaeaple. he cited hydroxyl groups on flat clay sur- 
faces bonding with hydroxyl rsdicals of polyners, which allowed metal 
ions to form insoluble phosphates. laMer (1967) suggested turbidity, 
subsideneo rate and floe filtration as aethods of ovaiueting dosta- 
billiation. He also suggested that a negative polymer would best 
destabUlte a negative eolloid. because many sites were producKi by 
polymer hydrolysis for bridging. 

Slrkner and Morgen (1368) measured particle size distribution 
during coagulation and found stronger Hoc was produced ns floe 
diameter increased. Th.y demonstrated the rat. controlling stop 
was partlclo agglomeration after coapilation, and that intenso agi- 
tation was responslblo for limited polymor sorption. OoIllMre and 
Horridgo (1972) invostignted flocculation of Chino clay using poly, 
acrylamides. They found that the maximum clarity was not coincident 
with the mixlHu filtration tatn as measured by the Kozony-Catmen 

sup«m8tant interface was the coneroUing flocculation paraaeter- 
Haiin ana Stiau (1968), scuil}lng tbe kinetics of alum hydrolysis 
for SiO, sols, dotomlnod chat there «ere three steps in the coagu- 
lation process: 1 ) foming polynuclear hydrolysis products; 2) the 

dal surface; and 5) the rote of particle transport. The rate lioi- 
ting step for SiO- coagulation was shown to he the race of particle 

the collison rate and the collison efficiency. 

also found that phosphates were reooved with Al*** In a cheoicol 
reaction, and the optianaa pH for the reaction was the sane as for 
optisBus bacterial flocculation. 

Stuaaa and Lee (1961) found that the rate of oaidation of ferrous 
I. They found an increase of one 

:h increasing pH. They found 
that silica retarded the hydrolysis of Fe**‘. The hydrolysis was 

with time. They suggested that Che solubility relationship may have 

pH unit near neutral pH resulted 
race. Schenk and Weber (1968) al 
dotion of ferrous iron increased 

tKe effects of teopereture on aqueous suspensions and concluded that 
sols, or Che nlua hydrolysis products mixed with cationic, anionic and 

•n C19S7) state. 

had a tendency to stahilire pH. He described sn 

that Bultivalenc hydrous oxides were sophoteric and that IT* and GH~ 
were primarily the potential dotoraining ions for suoh hydrous oxide 
precipitates. He also stated that aetal ions precipitated in the 

precipitates. Scums et al. (X9$71 formed polysilicates and elassi- 

nors and S] the mononuclear wall. They concluded from potential 
aeasurenents thot the interaction becwoen the anionic polymerlo 

overcome alectroscatic repulsion. They found optimal destablllta- 
tion occurred when a fraction of tbe oolloidal sorface area was 
covered and suggested that the nechanisa of destabilization for 

Stum C19S7) publisheij n hydrolysis model tor colloldel desta- 
blUiatlon tbit oMounted for hrldsio* sod olectroststlc effects. 

He postulated that a fraction of the total colloidal surface area 

aatbsaaticaUy using a Langalur isotherm by equating the amount of 
coagulant necessary to produce a certain fractional coverage to the 
sum of Che residual and sorbed coagulant. The fractionsl surface 
coverage necessary to destabilize colloidal sols could only be 

shoved from his model chat the required coagulant dosage to produce 
destablllaatioo could be independent of surface concentration or 
linearly dependent on surface concentration. In the Stieimi model 
metals first destablliaed colloids duo to sorption of the bydrolyaed 
cationio coagulants and restablllzod the colloids due to extensive 
sorption of the hydrolyzed metal coagulants, finally a precipits- 

lant become atteched to the colloidal surface, the coagualnt dosa 

If this occurred, the coagulant dose vas not a function of the 
colloidal surfaca area. If Che colloidal concentration vas high, 

sacion by sorption and final destabilization by precipitation mould 
be Indistinguishable. 

buffering were reaoved. A syntea with a aedt 
area require a atoichioDetrlc coagulant dose 

ta colloidal surface 
If low colloidal sur* 

required to coagulate bg precipitation. Scoiebionetry could be 

Kawaoura (1975] reported that CatOH)^ additions should be »ade 

reooval. Jeffcoat and Slngley (1975] found that CaCOHl^ addition 
prior to alua coagulation increased turbidity reaoval and recoosended 

Kenoah et el. (1907) measured alun floe site variations with 
kaolin, polymer and polyphosphate additions. They found kaolin and 
polymers increased floe site. Polyphosphates hindered floe foroui- 

guletlng high slkalinity waters with 
FeCOgis) may be precipitated instead 
O'Conner (1973) found arsenic was res 
chloride hydrolysis. They concluded 

(1961) found, when doing alknlimetric 

istead of Po(CH)j(s). Guilledge and 

adsorption was the removal 

base required to titrate the aluainua mixture vas not increased scoi. 
augsested Chat phosphate removal by alum coagulation resulted from a 

ohealcal reaction producing an alwinum hydroxy phosphate. 

difforent natural Haters was removed stoichiometTically by ferric 

removal did not produce a floe that had zero teta potential. Singley 
Unity had to be added before coagulation, ferric sulfate proved to 

packham [1P6S) studied coagulation of organic 

to be sinilar, because stoichiometric 
Here required to remove different 

empirical formula for such a reaction Has AlCOH),-, 

of these ooagulancs 

Packh^ fl9gS) did achieve en opclaun pH for color removal. Jobin ar 
Ghosh (1972) studied the azidation of ferrous Iron. They found that 

the ozidatiOD reaction. Schnicter (1971) found, at pH 2.S, that 

present. Monsrevlte et ol. (1975) conducted ezperlBonts on humic 

aluminum humic precipitation formed at a pi 
Al(OH),(s) precipitates. They suggested ti 

fulvic acids acted as anionic polyelectrolytes chat reacted chemically 

The reaction products formed a colloidal precipitate that could bo 
removed by settling after flocculation. The first step in humic and 

flocculation by cationic polyclectrolyte addition occurred. 

Luner and Donee (1970) determined that the color bodies present 

had a lOMcr avaraga Bolecular aeisht than the precipitated frectiana. 
Uiner at el, (197G) found, ulth flaaelva line troatpent, that enolic 

slum hydrolyses significantly at pH values encountered in line soften- 

cions. Gidsnoss and Slack C19S7] reduced the volune of sludge pro- 
duced in water toftening operations at Dayton, Ohio and Gainesville, 
Florida by bubbling CO2 into the sludge to dissolve HgC0i02> Sixty 
per cent of the MgCaH)^ was solobilised, bi 

because Hg{0Hl. existed os a gelatinous ct 
could accept a proton sore readily chan I 
CoCOj. This enabled the sludge volume tc 
suggested that not all the HgtOH)- could 

Thompson et al, {1972a) treated potable water samples im 1 

Hack (1997) concluded that 
10 carbonecion 

color and turbidity when C&CC^)2 “bs used to 
et ol. (1972a) proposed a potsbla water treat 

>m the sludge by car- 

bonatlon, and the CaCQH)^ was recovered froa the resaining CaCOj by 
recalcination. They proposed that sludge hsndling prohlaas associa- 
ted with conventional coagulation plants would be greatly reduced 

compared conventlonel coagulation systeas with the proposed magne- 
sium carbonate system. They also dnonstrated that as turbidity and 

GainesvillOi Florida. .1 

1) succeaafully ostended the magnesium car- 
0.1 mg/1 P, and significantly 

Slack (1974), in pilot plant studies at Melbourne, Florida, 
found evidence chat the color was released from Bagneslum sludge 

The potable water produced by using the recovered nagnosium as a 


to be acceptable. 

Predali end Cases investigated tete potential of msgne- 

sluB carbonates in electrol>^es and found OH ond H* to be the poten- 
tial detcraining ions for HgfOHl^. "Ibe MgCOiOjCsJ colloids had a 
teco lets potential at the sane pH for varying ionic strength aqueous 

■ust have been the source of the positive charge on the MgCOlOjCs) 
colloid. Zoganathan and Haier (1975) found that sand and haolinite 
colloids in a aolucioit of D.005 N HgClj had a positive seta potential 
for a pH of 10.5 or greater. They attributed the positive leta poten- 
tial to the increasing percentage of MgWl* relative to the total spe- 
cies of soluble BBgnesiuB. 

n Puln Mill effluents 

polynsrs could destabilize cc 
neutralits the electrical dot 

'0) desonatratod that vhil 
pulp waste, they would nc 

Dole (196S) studied sludges froa pulp bills ond found that floccula- 
tion kinetics varied with pH' for organic carbohydrate baae sludge, 
but not with an Inorganic priaary sludge froB a newsprint still. 

Their data indicated that aiost of the coUcidal bound water was 
interstital and was not che«ically held due to the solid-liquid 
Interfeco. Tho Nationel Council fot Air and Stroaai ImproveBent 
(1971) studied surface properties of hydrogels resulting frost trest- 

negetive colloids. They suggested that the 
the negative surface was nonstoichioaetrlc. 

Davis [1972). uslDg Cs(OH). co8|uUtion 'at Riceboro, Georgia 
to laaovo color froa a kraft aaste, found caltlua solubility decrea- 
sed as the Mdlua concomration Irm the dljostion operation was 
increased. Davis deaonstrated that orjanic carbon, color and caicioffl 
concentrations aftor Ca(0H)j treataont wore relatod to the initial 
sodiuia concentration of the waste. Berger (1964) found that a large 

that renoved 9M of original color from caustic bleach effluent. The 
Doaltar Limited Beaearch Center (1974) reported chat CnCOHlj coagula- 
tion was not effective for removing color from sulfito liquors. The 
Interstate Paper Corporation in fticeboro, Georgia used a smaller 
chemical dose of Ca(0H)j [1500-2500 mg/I) to resove in excess of 90% 

exceed the solubility product of Ca(CH)j end formed a precipitate. 
Othof and Eclenfelder [1974) studied color removal from three kraft 
mill effluents by separate coagulation with CaCOH)^. ferric sulfate 
and alias. They suggested that ferric sulfate was the better caagu- 
lant betause of lover coagulant dose and less voluminous sludge vol- 
umes. Could (1975) reported that the effluent from the cnuatlc 
extract stage of a kraft bleach plant, when treated with CalOHlj, 

color. Approximately 80% of the Ca[0H)j was recovetod in the sludge. 
Spruill (1975) found Ca(0H), treecment was very effective for reducing 
color in kraft wastes, but was ineffective for removing color from 
sulfite waste. Lestctynskl (1972) concluded that of the many pro- 
cesses proposed for color removal from kraft wastee, only Ca[C4i)2 


precipitation was feasible. Ksbeys el al. 
absorption of kroft Bill lignins on activated carbon to be very low. 

activated carbon in sorping lignin from graft Bill effluents. 

studied the Bcchanisni of color removal on activated carbon and found 
Dost color bodies eaisted in Che high ooleculer weight (15,000) 
range. TOC and color were not removod in o^ual proportions. They 
concluded that color reeoval by activated carbon was not e chenical 
process, but was due to sorption. Swanson et al. (1973) did a 

was no reDOval of uterial with BOlecular weights less than 40D. 
Hotorial with Dolecoiar weights greater than 5000 was completely 

lar weights ranging from 400 to 5000. Swanson (1973) suggested 
id a negative charge. 

UB, AlClj and PeCl^ as coagulants in 
c extraction waste and chlorinated 
rCi- and MCI. were capable of reaoving 

la chlorinated 

BC al, (1964) fractionated ai>enc sulfite vasce liquors by gel 

rides and weak organic acids at pH 4 ware present in the lower a 
cular weight range. Praetions above a aoleeular weight of 40. 0( 

Soith and Chriatuan (1969) treated kraft and sulfite waste 

906 of the initial color in the Kraft waste. Treataent of the s 
fite waste b/ PeCl^ reduced the organic carbon 606, but increase 
the color of the sulfite waste. Alum reduced the color of the s 
fite waste 676. Salth and Christsan (1969) proposed that the kt 
waste had sulhydiyl groups on lignin chains and chat these group 
foraed insoluble sulfides during coagulation. The sulfite waste 
hud sulfonate groups in thn lignin chain which acted as strong a 
and foraed h/drolfsis products. The aechaoiaa for color removel 
Che kraft waste was suggested to be a chenlcal reaction, vhcraas 
nechanisQ In the sulfite waste was soggestcd to be sorption on A 

IS acconpaniod by the fometion of a 
than the original CalCH) floc. A 2 


3-1.1 Synthetic Waste Solgtions 

fran a NSSC plant lUgestor after the cooking operation had been coai- 
pleted. Thia liquor contained the dissolved constituents of the 
wood. It was referred to as "sulfite waste liquor," which can be 

eted by Head Corporation and Sunoco Products Conpony respectively. 
ie Head Corporation supplied soldiun base spent sulfito waste liquors 
Id asBOntua base spent sulfite waste liquors that were used as a 
isto source. The Sunoco Products Coapany supplied a sodiusi base 
lont sulfite waste liquor which was also used as o waste source. 

There are different processes and nnny different types of hard- 

lis, it was decided at the beginning of this research to deteralne 

color of SOOQ Pt-Co color emits, o dilution ratio of SO/l to 100/i 

of SO ns/Bl as M|** was Dade in order to atninize the volume of 
coagulant feed dosed in the process. This was achieved by dlssol- 
ving 532.6 grams of MgSQ^'PH^O in a liter of dUtilled-deionized 

Caloiua hydroxide and sodiua hydroxide were used for pH adjusc- 
aary, sulfuric acid and hydrochloric acid wore used to adjust Che 

Cationic, anionic end nonionic polyuers were prepared free 
cooBBerciel lit^uids and powders supplied by American Cyanamid Conpony. 
The polymers were nado frc« polyacryla«ide and axine bases. Stook 

supplied by Aaerlcan Cyanamid (N-478] was T.equired for stock prepara- 

for analysis. The readout was registered on a 0 to 100 scale ai 

The pH of tho sample was then regulated to 7-6 before the aac 

OB a standard curve relating color to abaorbance. If the aanple bad 
too great a color to be directly poasured. the sanple ues diluted 
after filtration. 

A standard curve vas prepared by dissolving 1.246 grans potss- 
sluB ebloroplatinate, {equivalent CO O.SOO g metallic plati- 

nua) and one grcia cryscallilod oobaltous chloride. CoCl2*6H20 (equi- 
valent to 0.2S grams netallic cobalt) in distilled veter vith 100 mi 
concentrated HCl. This solution vas diluted to I liter vith dis- 
tilled vater. This stock solution was defined as having a standard 

Color {218g.4)(absorbance) ■ 4.4 (3-1) 

Sludge incineration was determined in a Thermodyne furnace, 
Hodcl P-Ai7S0. Sludge samples were dried at 10S°C and filtered 
through a Buchner funnel on a Whatman no. 40 ashloss filter before 
incineration in the furnace. Incineration temperatures were varied 

minutes to UO minutes. 

re dosed simultaneously Tt 
ice at 100 rpa for throe s 

r Test Machine. Cheai- 
r beakers. Rapid miming 
Florida Jar Tester was 

le highly colored wi 

a stirring rate of 100 nts the bui> 

brolien by the stirrer 

ou Bln and Che rapid Bin cycles Co 

allowed CO settle for 30 BinuCee before sajeples were taken for analy> 
sis. If the coegalatins Bixture had not developed a clear superna- 
tant, e saaple was taken end filtered through a no. 40 Whatman fil- 
ter in order to slaplify the required filtering step through the 

The G levels of the rapid Bin and the slow Bin cycles ware 110 
sec'* and 30 sec'* respectively. The Bining level in Che floccula- 
tion stage had a Gt vsliie of 27.000, which was appronintatoly the low 
end of Che range spaelfiad in Weste TreaCBent Plant Design (1971) . 

Metal analyses were deteiBlned on a Varlan Techtron Ht 

Bade at a wavalongth of 422.7 nanoeeters. All saatples were filtered 
through a 0.00 aicron Millipore filter and treated with 1 ml of 17k 

siua volues were oeasured. 

3-2.6 Mobility Measurements 

in conjunction 

Riddick call and a see 
uaod In the Riddick CE 

> on c Comlns Model 12 expanded 
•silver chloride reference elec- 

3-2.8 Settling Tests 

ted cylinder- One liter of waste woe coagulated in a jar and lane- . 
dlately transferred to a graduated cylinder where the height of the 
sludge-supemetant interface was recorded. The following fomuia 
was used to calculate the Sludge VoluBO Index, SVl: 

SVI • ml settled sludge x 1.000 
Bg/1 suspended solids 

3-2.9 Solids Analysin 

All suspended solids analyses wore deterBined on aomplos that 
were filtered through a no. 40 Whatman filter and dried at 103 C fo 
one hour. Nonvolatile and volatile solids wore by filtering the 

one hour, and recording the weight. The sample was then ignited 
SS0®C for 80 minutes, after which it was weighed to determine 

)1 and 0.1 N HjSO, and NeOf 

te strength of the soAples. 

voluoe of waste titrated 
One minute was allowed for pK stabill- 
.rant was added to the .sample. A Teflon 

I during titration. 

. Coopilatlon Emeriaents 

>n and moasuring tl 
described. Ibe next step was regulation of 
with CaCmij or NaGH. The coagulant was thi 

of settling. Organic cc 
determined Immediately after coagulation. 

.e solution pH 

Id immediately following cc 

Following coagulstion the resulting sludge vas filtered through 

then dried at 103®C for one hour. The dry 
SSO^C, end the resulting nonvolatile solids 

the oxidiied Dagnesiuo. Two 40 liter volumes 
in order to produce a large quantity of sludge, 
chon heated until they achieved a constant weight 

3-3.3 Coagulant gecyclo 

The magnesium was recycled to determine the effectiveness of 

removal process. Tho Bethod of recycling the magnesium consisted of 

SSO^C. The resulting nonvolatile solids were carbonated for 45 
minutes with e lOh COj-POl air gaseous mirture. The recovered megne- 

recyeling the sugnesium determined the effectiveness of the remaining 


Ca(0H). and NaOH bocause they vore inexpensive and eooDBarcially 
the Qdnimia pH and dose chat resulted in a 90b reduction of the ini- 

A three step technique wes used to determine the coagulation 
pH and coagulant dose for the color reoaoval process. First the coagu- 

coagulation pH. Finally the stability of the coagulation pH was 
verified by repeating the first stop for a nagneslun dose other thsn 

!s fron which the coagulation 

wch curve is indicated un the figures. The miniBiBB pH at uhich 901 
celor re»jyal »as achieved vts 10.6. The final color was dependent 

of the |W control agent. At fdi 10 when Ca(0H}j was used, the final 
color of the waste was increased approsleately 30% sore then thn lal- 

reaoual was obtained free pH 10.6 to [dl 11-4 using Ca(»l)j or NoOH. 

The degree of color renovsl decreases past 11.4 when MaOH was used. 

These figures show that a 90T reduction of the original color 

BS first reached at pH 10.6 for oach color. This indi- 
coagulaclon pH was independent of the variebility in 
r. Froa the data presented in Figures 4.1 and 4.1 it 
the coagulation pH was 10.6. 

4-1.2 Cosaulant Dose 

A HS5C waste with a color of 2500 or 5000 was coegulated at pH 
10.6 with a varying vagnosiuB dose- These data are pmaented in Pig- 

nesiua coagulation of the MSSC waste with either pH control agonC 

for pH control. The nagnesiun dose 
for pH control. The reqoired aagnss 

color of 5000 was 200 og/1 when CafCHlj was used for pH control, and 
was 400 ig/1 idien NaCH was used for pH control. The rOHUlrod coagu- 
lant dose was directly proportional to the color of the NSSC waste. 

Ihe SABe color roduction was achieved whea CaCOK)^ or NaGH was 
used for pH control. As is indicated in olcher Figure 4.3 or 4.4, a 

sot less tdien CsEQH)2 was used to control pK. 

4-1.3 Variation of Coagulation pH with Coerulanr Dose 

All KS5C waste solutions were asde by diluting a stored hSSC 

4.S shows for a color of SOQO, the optiBum color reaoval again occur- 
to control pH. This was the same pK at which Bnsinua color renoval 
junction with CaCOH),. It was therefore concluded that the coagula- 

was 40,000 mg/1 and 30,000 Bg/1 respectively. The amount of sodiiA 

The eddltlonol sodiu* added W the vasto to adjust fdt «aa not s 
eipiiHeant itictoaee in sodlim conoenttaelon. and did not esteneively 
foTB any conplctes. ' 

approiimatoly 40 mg/l as Ca“- When Ce(0H)j was used to adjust pH 
the calelm concoBtration In the waste wss Increased to 600 m*/l. 

This was sipiifioant becsuse tha ealciun coneentration Increased 
and pmhaMy did e«enslvoIy comploa the organic cooipounds. 

Pljuros 4,5 and 4.4 show that the nagncsiuoi required to remva 

rather than NaOH. Both calciuai and magnesiuo are divalent ions and 

When Ca(0HJ2 was added to control pH, Ca** conplexed nany orgonics 
that Mg** would have nonaally complesed in the absence of the added 

A negnosiuB dose of 100 ng/1 renoved 00% of the color fron a 
NSSC waste with a color of 2500 whan Ca(OH)j was used for pH control. 
A negnesiun dose of 200 ng/1 was necessary for 90% color renovsl 
when NoOH was used to control The conplaning ability of the 
caiciun ion was responsiblo for a 50% rednetion in the coagulant 
dose. Since NaOH Is more enpenslvo than CaCODj and does not reduce 
the cosgulant dose, Ca(OH), wns chosen as the pH control agent. 

The magnesium remaining in the treated waste, t 

MagoosiliD Iji solution raaalsisd approsiisatel/ constant past 11. Coagu- 

Mstad In the treated effluent. The CaCffi)2 dose was Increased 
approximately SO mg/1 to raise the coagulation pH to 11. The ugne- 

Increasing the coagnlation pK 

tion pH was increased. The per cent color removed was significantly 
a reduction of C.S pH units would not significantly affect color reaiovs 

i HagnesiuB and Ca(Cf{l2 ' 

IS a Function of In 

The waste effluent from e semicheaiicel neutral aulfite plant 
varies in color Intensity. Seeause of this veriability, the amount 
of Ce(0H). and magnesium te remove SOS of the Inltiei NSSC color was 
deCerminod as a function of the initial color of the waste. The 
Ce(CM)j tenulred io presented in Figure 4.7. The magnosium require- 
ment is presented in Figure 4.3. Both the CaCOK)- end the mogneslun 
roquireaents were directly dependent on the initial color of the 

waste Characteristics 

4-2.1 Untreated Waste Titration Curves 

llie acid stren^h ef the untreated hSSC waste was determined 
b/ titrating SO ■! saaples with varying calors with 1-0 N H^SO-i- The 
acid strength of the waste was defined as the milliequivalents of 

these titmtions are presented in Pigures 4.9 through 4.13 far NSSC 

waste colors of 2,500 to 40,000. The NSSC waste was obtained freat 
the Sunoco Products Corporation in Hartsvllle, South Carolioa and is 
denoted as sodium base Sunoco NSSC waste in Figures 4,9 through 4.13. 
As the color of Che NSSC waste was increased, the acid strength of the 
NSSC waste also increased. This indicated that color was acidic, and 

Iheso functional groups had pK values in the range of carborylic 
acids end phenols or enols. The equilibrium constants for the NSSC 

The pE values were idenclfiod by locating the inflection points on 
the titration curves. Approximately 6db of the data points are not 

areas, both of •diich were identified by slight humps on the titration 
curves. These pE values are approximately 4,s and 9.8. They differed 
by four orders of magnitude, which was a Urge enough separation Co 
allow graphical approximacion of pE values. 

ns. 4.9 

NSSC •!>$« 

Flfl. «.l 

wttti cdor fo SC^OOO 

ng. 4.1 


I Coaparison of Untraated at 

:e Tttraiion O 

Tha acid strength of the untreated N5SC voate uoa coaipared with 
the acid strength of the created NSSC waste. This was detoo In order 

Pig. 4.|. 

A. tclmt of 200 mis of NSSC waste 

solution to 0 pH of 12,0. The tltratii 
presented lo Figure 4,15. la the fim 

The solution was slightly buffered by 1 

all pH values greater than or equal to 
tioo carves, magnesluii was acting as a 

was dosed with 80 ng C400 og/1) 
n curve for this experiment Is 

. was accomplished when tl 

The acids we 

E ionised oj 

SHght hjffar^g 

>r resi<AiaU si 

IT end Organic Carbon RoslduaU After 
A NSSC vaste with an initial color of 2.SOO waa coa|ulacad with 
was varied free 10 to U.S usln^ Ca(OII}2 to control pK. The oa^nc* 

N5dC waste was coagulated at pH 11. S. The residual color at this 
ony pH froa 10.6 to 11.2, approxiatacely 34A of tho total organic eai 

coigulation at pH II. S, the color wsa Toduced by only 1,5 Pt-Co 
orsanie carbon in the woste contributed equally to the waste color. 

The residual color increased 324 when the coagulation eaperi- 
■ent was acteapted at pH 10.0. The aagneslum residual curve in 
Figure 4.17 shows that no BognesiUB precipitated out of solution at 
pH 10. All of Che aagnesiuB wos therefore available to form che- 
lates with the HSSC waste. Caleiun ions causing increases in the 

Dence (15711, Color increasing chelates Fomod by aagnesiua and 
quinones have been reported by Day and Underwood (19671, Arosiatlc 
quioones are an integral part of hosie lignin structure, and lignin 
is responsible for color in pulp waste. Tho color increase at pH 
10 was probably due to tho chelation of lignin building units, pessi- 

at a final pH of 10.0. The total amount of aagnesiua available for 
coagulation was the nognasium dosage and the mogneslua present in the 

161 ag/1 magnesium. When magnesium precipitation begsn, a cortespon- 
dlag drop in color intensity was observed, as was a corresponding 
drop in organic carbon. Tho concencrations of asgneaium, organic 

decTOAses in ngnesimn, color and organic carbon occurred siBUlta- 

cculd be a chaical reaction between Che color producing organic coa- 

ohelation and praoipltation of a nagnesiua organic color-body coaplea. 
A aocond possibility could be the adsorption of the chelated organics 

Figures a. 10 and 4,17 could confers to either of these postulated 

CaCCH)^ was used as tho pH control aj 
ks the Cb(OI 02 dissolved, the pH and celci 


nesiuB concentrations can be visualized as a reaction between the 
dissolved CaCOHlj end Hg**, Such a reeetion is shewn in Equetion 4-1 

deterained after coasulation vith a vatrios Bagnesium data. In these 
ex^erinents, the pH was held constant at 10.6- The pH controls used 
uere NoOfl and CaCOHlj. The date froa these experiaents are presented 
in Figures 4.U and 4-l», In those figures the residual color, organic 
carhon, and soluble aagneslua arc plotted as functions of the total 
Bllllaoles of aagnesiuB availehlo. The total mllliaolos of aagnesiun 
available consists of the Initial aagnesiian plus the coagulant dose. 

appresiaately half that of the initial color of t 
Figure 4.19. The scales in Figure 4.18 are one-1 
Figure 4.19. This »aa done so the tuo figures cc 
pared without being aisleading. 

a decrease in color renoval, The beginning of f1 
tified by the dashed lines in Figures 4.U and 4. 
identified when the floe becaae visible to the m 

the totel aaount of available Begnesium was 2.19 aillisoles. This 
point was not reached with NaOH until the total ailllaoles of Hg**^ 

available vere 9.46, If 

treated Mith Ca(0H)2. If the ao^esiun dese aaa directly propor- 
cioaal CO the IMtial color, then eacrapolating the Cb( 0H]2 treat- 

is found that 5. 08 nM/1 iBore of eagnesiun 
^ UBlog NaOH. This difference vas due to 
m CaC0IQ2 vas used to control pH. 

m ions from CaC0H)2 cosploxed sobb 

the NsOH treatnent . Il 

Ploc fomintiDn did not coo 

■agnesium had Ca{0H]2 not 

have been identified in the color reBOVol process. First, chelation 

before color reiDOval xcurred. The magnesium chelates were partly 
reduced by using Ca(OH)2 to adjust pH. After the chelation demand 

A possible fflMb&nlsis of color resoval was adsorpElon of Che color 

pound, but would hove becone ecceched to the floe bf Vsn der Kaals 
farces or hydrogen banding. If aagnesluB hydroxide floe were fomed, 

reooved froa eolutioo. If eji insoluble precipitate formed that was 

ions, the soles of magnesiuo ions removed divided into Che moles of 
hydroxide ions resoved would be less then two. 

Ihe poles of Bognesius removed divided into the soles of hydroxide 
resoved is presented as a ratio in Figure 4.20. There are three dif- 

Ca[ffll)j was used to control pH for color removal fros NSSC wastes with 

hydroxide resoved were found by difference. First the moles of 

amount was subtrected from the moles of hydroxides required to raise 
the pH to 10.6 after Che aagnesim dose was added. The differenoe 
was cha hydroxide demand of the magnesium used to coagulate the 
color, and was reprosented as In Figure 4.20. If.KaDH was 



)les of Fi/droxlde vore o«asur»d fton the 
I NatM solution. If CaCOH); •'as used to 
.0 oalciiiB concentration before end after 

direct addition of a ; 

coagulation was aoasured. The calcium incroaae was doubled to deter- 
mine the moles of hydroxide required in the color removal process. 

coagulation was 10.6. Both CoCOtOj and HaOH wore used to control' 
pH. All of the data points in Figure 4.20 represent sono degree of 

Hhen KaOH instead of CaCOHJg was used to control pH, a greater 

curves for itastos of equal initial color in Figure 4.20. 

The initial points on each of the curves in Figure 4,20 repre- 

matlon increased, the curves eventually stabilised at l.S, At the 
low magnesluiS dosos used initially this rstio was not stable because 
of the chelation demand of the waste. Once this deieend was exceeded 

duce an electrically neutralised compound. Another anion had to con- 
tribute one-quarter of the total negative charge for the precipitate 
to be electrically neutral. Color bodies are negatively charged and 
color was removed ss magneslua ions were precipitated. IF the color 

:h the etgnesium and 

ces that an inaeluble precipitate 
: 1R~ where R~ represents the 

negatively charged eclar bodies waul 

The fgrmotian of en insoluble precipitate is further supported 
becsne eoBeshed in a HgC(M)2 precipitate, the overall OH/Mg ratio 

because Bostly Mg{CH)2 would be fomed after the eheletlaa demand of 

does not vary after becoaing stabilized at l,S. Proa the data pre- 
sented, it was concluded that a chemical reaction was Che Beehaniso 

a hraft waste at a pulp plant in Ricoboro, Georgia. Dissolved CaCOHJg, 

of CatOKl2 with reference to a NSSC waste. The residual color and pH 

chel&tion of esUiia iaiu \ 
muicuv CbCGH ^2 

re detorained in a NSSC vcste efter CAfOH)^ sddition. These results 

) BS/1> *rith a resultant tdl of 12 and 

pH 11 the use of CaCOlQj alone slightly increased the residual color 
of the waste. This Is shown in Figure 4.21. These results indicate 

I Settling of Coagulated hi 

4-4.1 Purpose of Settling Tests 

sludge and gain soee knowledge of the factors governing the settling 

The settling tests wore conducted as described in Section 3-2, S 
of Chapter 3. Settling tests wore perfemed on .the coaguleted waste: 
snd on pelyoer treated eoegulated waates. Cationic, nonionic, and 
anionic polyiaors woro nsod as setcUng aids in the tests. All of 
the polyners tested were supplied by the Aaierican Cyanaoid Ccepany. 

4-4.2 Sludeo Settlsability 
The type, functional jrot 

Table 4>2. The Sludge Volupe Index is 
as a fuection of polyoer dose and polj 
The 5V1 of the raw sludge was 3s; 

le settling tests is presented in 

I, coppletely dispersed and 

if eosapression during tl 

of the nonionlc polyiaer s 

1 of a Donionic polyaer s^s used as a a 
ced to 17g, The physical appearance of 

IS effective in consolidating the floe. 

Hj” » 550 »g/l, Ca[0H)j ■ 1500 mg/1, pH = 10.8 

'e in connoUdBtlng ti 

The saaller pnlymor created repuisive 

charge. The nonioeic polymer waa larger and r 
floe setcleability reeulced from the polyner-1 

settling a 

spheres approximately 0.5 hb in diaaieter. The polyacrylic acid 

a significant reduction in SVI hut an excessive polymer dose vns 
required. The available surface area for floe intemctlDn was oucl 
less vhen the activated spheres of polyacylamide were foamed. The 

coagulated nixtiire as clear liquors and were cosipletely soluble. 

The interaction between the anionic poi/ncrylanide end the floe 
was quite rapid. The rapid interaction is shown in figure 4.22, The 
sludge Interfoce is preaonced es n function of settling tine. The 
solids concentration in the sludge was increosod cpproxiantely seven- 
fold duo to Che addition of 3.0 ag/1 of a 51 hydrolyzed polyacryla- 

The negatively charged polyacryloaide was the nost effective 
settling oid. A high degree of negative charge was not required on 
the polyier. This wes shown by the identicil effectiveness of 537A 
and g35A. A polyaer is negatively charged by hydrolysis. The grea- 
ter the degree of hydrolysis on the polymer, the greeter the polyaer 
charge. The 837A polyner was St hydrolyied, and the S35A polyaer 
was 2St hydrolyzed. A St hydrolyzed poiyecrylaBide means that 95 out 
of every 100 Bononer unite are uncharged acrylamide groups; the 
remaining 5 monomer units, ere negatlvoly charged acrylic groups. 

Charged polyacrylamide; 

Fig 4.82 Sludge seTlIing velocity 

K/drolysi5. Sorae degr«o of interaction be 

lOt ao effective oe either 

destebilitation cf colloids. One area deals wj 

Due Brea deals with the collcidal st 

then can agglooerate and settle. Kovever, there are aany iMssiblll- 
betveen colIoid-coaguUnt interactions, Lamer et al, Ci9b7) have 

eollold&l descabillsation to occur by this Dodel, the polyaer dose 

dote or by shearini the polyaer with coo high a cdxlng energy- There 
are aany inatences in wastovater treataent where negatively charged 
collotda are destabiiiied by anionic polymera. This phenoaena 
can be eapUlned by an interaction between the functional groups and 
the colloids, os in the bridging aodel. 

The electrctphorellc aobUlty was measured on the floe particles 
to determine if the settleabllity of the floe particles Inereased as 
Che floe charge decreased. The Heleholta-Saoluchowshi (H>S) formula 
was used CO decemine the eleccropboretic mobility. Riddick (1974) 
specified Che applicability of different aeta potential fortaglas 
based on normslity of suspending solution and particle diameter. He 
recoeaiended the Helmholti-Smoluchowski formula to measure the electro- 
phoretic mobility of aay particle suspended in a l.ON solution whose 
diameter was 0.8 microns or greater. The floe particles produced in 

The K-S formula for determining seta potential is as follows: 

IT - 113,OOOC9t/Oj)B4 (4-2) 

ZP ■ Zero potentinl in millivolts 

EM ' Electrophoretic mobiiity in microns cm/$ec volt 

function of polymer dose for n nonlonie end anionic polymer. It 
will be shomn later that e negative potential occurred on the floe 

eloctrostatic reduction was the major noehaniam of enhanced settling 
of the negatively charged floe, then the cationic polymer would have 
been the nest effective settling aid. As Tabic d-3 shows, Che cat- 
ionic polyuttr stabilitod the floe and severely hindered settling. 
Conversely, the anionic polymer was seen to bo on effective settling 
aid. Restabilitation of the floe was not achieved at the polymer 
doses tested. The seta potential was observed to increase from -13 av 

It did not approach tero although tho SVt of the sludge changed from 
332 CO 83. Tho total change in ZP os settleability increased indi- 

mechanlsm for floe destebiUzatlon. The controlling mechanism was 

and N5SC waste, electrophoretic swbilities were determined on magne- 

Ihc data for these experiments are presented in Table 4-4. The teta 
potential as e function of pK is graphed in Figure 4. 

the tap water was positive, and increased with increasing pH. Tho 














5 . 






pestesi change in seta potential with raapact to pH occurred at pH 

to the fomation of The equiUhriua concentrarlons of 

and M|** are prcsanccd in Ptguro 4.26 froo the equilibriiin 

Sjp " CM8**H0H'1 ■ 10 ' (4-3) 

pH. The ratio of the aingolaxlf h/dxoayUtod specioa MgCOH)** to Hg** 
tration of Mg(Dtl) decreased. As the concentration of Hg[0H) and 

4.26. one can suggest the flop was positlvel/ charged in tap water 
hydroa/Iated species. 

coogulaeion of the NSSC vasts changed significantly Co pH 10.7 and pasi 


IB negatively charged. The cc 
me negative tete potential 

loida aialntained a relatively cc 
through the range of naxlBiaB color reaioval, pH 10.6 to 11.5. Hhen 
Che ^ vea raised to 12.5, the teta potential increased to aero as a 

the nagnesiun floe. The ability of the hydroaidc ion to successfully 

reduced when CaCOH). was used to control pH due Co the coeiplerlng 
ability of the cslciun ion. 

1 Hogoesiua Recevery and Recycle 

a recovery process cl 

inclneratod aasnesluB sludge. 

A color rolOBse experlDont wi 

Cft(CH).. The final pB of the coagulated waate was U.O, and 

measured, and then the pK was raised to 11. Q and the color was again 
measured. This oscillation of pH completely dissolved and reformed 
the sludge, hhen the 'sludge was dissolved, the color returned to 

These data are presented in Figure d.27. The reversibility presen- 

with only carbonation would be feasible if the precipitated color 
remained on the CaCO^ floe during ci 

4-S.3 Color-Cation 
designed to determine if CsCO^ precipitation in situ with the 

In the CeCOr precipitation exporieenc, and CaCCH^^ 

CaCO^ preclpiCBted per liter ranged from 0-2 to 260 milHoolas- 
Since l«2®j ah' source materials, the amount of 

checked by Che carbonate difference before and after coagulation. 

The carbonate difference was meaeured by determining the total inor- 

chango as a function of the aillijnoles of CaCO^ precipitated is pre- 
sented in Figure 4.28. The dose data and the change in TOC concen- 

cium was not available from i 
froa the Ca(OH)j dose for s i 
approximately a db color removal based on tl 

:o be removed from the magnesium floe b] 

s significant 

nBEnesiuB precipitate that was not a hydrol/sis product. Mgp2 

Different Jars of N55C waste with an initial coier of 29S0 were 

All jars were in e state of MgFj supsrsacuratf on . The general defi- 
nition of the suporsatuntion ratio (3) is 3 ■ where Q is 

the Ion product, K is the equilibriua constant, and n is the nuober 

allowed to stand for 24 hours before aanples were taken. The floe 
formed was very small and not nearly as voluminous as the nagnesium 
floe produced at high [dl. The residual color as a function of the 

lha ioitlal color of the N5SC waste was reduced SSt when the 

point all of the nagnesium dose, 7.3 nmole, had been precipitated 
from solution. In the MgF^ exporinent, the precipitation of 1 asaole 
of magnesium removed 231 Pt-Co color units. If tdoto magoesiuo hod 

might have been larger. However, a significant nagnesium-color inter- 

involving magnesium precipitetion. 

of one (TBSole of U|ne5iua rewved approxifiacely 360 color units fron 
solution. Comparison of the ratios of color removad per mole of 

tive than calciuo. The, ■agnesiuB sludge produced at high pH was aore 

orfi given In TnOle 4,6. TorOMgiOHk 
(^UgCOj, (SlUgCOj' JhjO, one 
®Mg4(C03)3(OH)2'3H20 which ore 

Fl9 4.32 ftidoflMnance diagram for log Mg** 

MgCO^'dHgO Is only stable tnermo- 
dynomlcoly or high PgQ^* HgCOj3HgO, 
Mg^lCOj)j(OHIj'3MjO and HgtOW, 

t Hg** concentration of 10“ 

, The activity ratio and solobi 
1 total alkalinity of 10 “* N. 

.0 and solubility dlagrus in Fig 

10 controlling species was «g^(C0j)jC0Hl 

practical significance of these dlagraas was 
controlling the solubility of Hg*“. For a Cy 

-JHjOCs). Below pH 8. 

If MgCOj-JHjOts) i: 

olkelinlty. The 
idontlfy the species 

in Figoro a. 31. For e C|. of 10“^ M an equilibrium point was reached 
at ^1 7.2. This is identified by the intersection of the equilibrium 
lines in the activity diagram of HCOj end HgCOj'SHjOCs) . The bicar- 
bonste ions from the solubilising of CO, gas are in equilibrium with 
the solid HgCOj'SHjOCs). 
proton fro. HjCOj*. This 

MgCOj'SHjaCs) ♦ H* • 

1 . 10“-®“ • (Mg**: 

IS represented by the following reaction 

PH • 


It wts interesting to note fros Che activit)' ratio diagram in Figure 
C.30 Chat Che species HgCO^Cs) never does exist at equUihriuD, 

species was MgCO^'SHjOCs). 

a 10% CO2 gaseous aixTure in an aqueous solution, 3.g9 x 10'^ aols of 

94% of it in the HCO^' fon. The naxiaum calculated oagneslua solu> 
hllitf considering KgC03*3H20ts] os the doainant species was 663 ag/l. 

Hg** dissolution fron HgCOHi^ts] was kinecicallF favored as compared 
could exist. The presence of organic acids in the 

color, because the organics wt 

4-5-S Sludgo Incinennion 

Tlie colored sludge Chet nee Co be incinerated was prepared by 
creating two soparace 40 liter volu&es of NS5C waste with 350 Bg/1 
Hg** and 1250 ng/l of CatOK),. An everage of four separate analyaas 
of NSSC waste and sludge is presented in Table 4-7. All sludge 
samples were dried at I03'’c until constant weights were obtained 

ag/1. The magacsiun 
1082 mg of solids re: 

as Che incineration of the sludge pro- 

l and 362 mg of nagnesium were accounted 

the sludge through a no. 40 dhatmnn 
Me very small particles of solid aagne- 

hably due to the filtoriog o 

pore filtering apparatus usd 
A cbemical representation 

jB - >60 t COj • other gases ed- 

it varying times and temperatures was iapleaenced. 

I 5SO as/I OF Mg** 
WmAt COLOR 01 

teaperaCure required to reaove the color bodies from the Bagneeiu* 
solids, however, a lialcing factor influencing Che rccoverjr of the 

■nm specific gravity of ■agnesium varied considerably with the tem- 
perature and time of incineration. This is presented in Table 4-0. 

A decrease in reactivity of HgO was paralleled by an increase in the 
density of MgO resulting from increasing calcination temperatures. 
According to Harper [1967], the freshly formed MgO had a high surface 

temperature, dead-burned magnesia resulted. The dead-burned magnesia 

900^C and was very unreactlve. Harper [1967) also found MgO prepared 
in the range of 400-900^0, called caustic burned magnesia, was readily 



4-5.6 MBgnesiuB Rocovary 

Hegnosium recovery by recarbonaTlon waa salected as the recovery 
tag the nagneaiiB with HjSD^ was 50.24/1000 gal. It was aoncludad. 

la slurry was agltatod by 



Because log K for HgCOj'BH^Ots) is nesscive for th® racovery 

raactioni a dacrease in tcflparature would docroasa tha aaounc of asg 

Kg** and fCOj at equilibrlua. The taoperature offsets on the ra 
of KgCOj'SK^Ots] precipitation are illustrated in Figure 4.34 in 

Fig. 4.34 Fr4Ci»niMcin of MgC0j'3H;0 b> ocrotlon al vorioui tenigarolijres 
S»utiM i aiiclc,A.P, £F« t»l2ISO HMZ,Sapt.,l974. ''Fu« Sc4l« StiMln 9> lha 

Alobaso. The effect of increeeirig CO^ pressure 
The predoaietnee diaisram presented c 

experiment, the time ef esrbenation 

Figure 4.32. The meg- 
.nutes. The pH resdings 

The plot of pH versi 
thst an equilibrium pH w 

equlUbrium pH was 7.6 a 
the loboratcrry studies. 

carbonatien ti 

' . The theoretically predicted 

le equilibrlioe conditlen could 

for the difference 
e magnesium carbonate compound in the 

^ supports Che formation ot 

In Figure 4.36 the magnesium eoacentratian after the carbenation 
of nonvolatile solids is shown as a function of .carbenation cine. 

nation as n function of Che nonvolatile solids concentration. 

TABU 4-10 


Rj. 4,3 

% Mogntsluffl ttcovary as 

The aesnesiua coneentretien efter caihoQBtion was never limited 
Ly the foroetisn ef o solid isagiiBsiiim compound for Che carhonstion 

trocion hail bocome conetant during carbonation and some atognaslum 
etill remained in the nonvolatile solids- As shovn In Figure 4.36. 
the magnesium ccncencretion was elways increasing during the carbo- 
nation process if there was any aagnesium remaining in the nonvolatile 

The maalmun theoretical magnesium oonoontration at equilibrium 
is 866 ag/1 if H|COj-3HjOts) was the controlling solid phase. This 
was Che controlling solid phase predicted by the theoreticai 
Kg'^-COj-HjO system presented in Figure 4.32. The controlling solid 
phase was not determined by these experiments. It might have been 

was not tttCOj-SHgOCs] 
rate of MgO(s) or MgCOHljts) 
nucleation of HgCO- 

soUd phase 

fcinecicolly favored with respect to 
lH_0(s). Another possibility was chat other 
le incinerated solids were involved ii 

4-5.7 Wssaiesiua flense 

IncinftiAtion and csrbonatlon procesios van used to recover tho 

could be rucceesfuUy reused Id T}to coagulation process. 

The BagneslLaa was recycled twice. Following each use, Che mag- 

were not dissolved during carbonaclon. These reeaining solids were 

on the soluble Dagnesiuia in Che carboneted liquor and the coagulant 
dose relationship shown in Figure 4.B. The date fron these eagne- 

che Bagnoslua contained 99.6b of the initial nagnesiusi. However, 
free the first use was carbonated. The color of this solution was 

wu probably doe to the inccenplete conbustion of the color bodies 

77b noMolatlle solids after the first majnesiiss use- The sversge 
nonvolatile solids reduction In the previous axperiaents reported 
In Toble »-9 for 5S0“C and 15 Binutes was 551. The solids loading 
rate for the incineration process was never optlnised. 

after it bod been used in the color rei^al process. Greater than 
90b of the initial color was reooved with each of tlie three eagne- 

rocovored oagnesiuts and the undissolved solids removed as much color 

recovered and recycled in the llme-aiagnealun tolor removal process 
sftor the incinorstion ond recovery processes. 

A second set of magnesium reuse experiments was conduersd to 
determine the effectiveness of the recycled solids that were not 
dissolved in the carhonation process. The rocovored magneeium solu- 
tion was filtered through s O-dO micron Hillipore filter to r*ovs 
Che solids. Tho per cent of color raoaved by the used magnesium 

ir chan did an equivalent ai 
re used separately in the ci 


reused Dssnesium in tKe color rcaoval process. These soHds proh> 
ably provided nucleating surfaces for the forming solids phase. 


led froa the optiteue dose equation 
I 4,S. TliB Ca(OH)2 requireacnt ve 

presented earlier is 

*c ■ 

Ag = Area required for clarification - ft^ 
V,j a Kindered settling velocir/ - fpm 

*t ■ Qtu/«o (5-5) 

Aj a Area required for thickening - ft^ 

a Initial height of sludge interface - ft 

HoC<, a Hj.C^ (5-4) 




The settlini velocity was decersiined to ha 0.187 ft/oio from 
le hiadered portion of the settling curve in Figure S.l. The ini- 
Lai and final heights of the sludge interface wore L.12 ft end 0.20 

The area required for thickening the sludge to a final concon- 
ition and therefore controlled Che area for settlieg. A design area 

duced during ti 

would he 15,012 Ibs/mgal. If carbonation were used to recover the 
CeCOg during coagulation. This would increase the ness of solids 
adjust the waste to pH 11, 16.0 tfd of Ca** was available to preclpi- 

would have to equal 0.22. The total carbonate concentretion in Che 
recycle streasi was not dotemined. For dosign purposes, all of the 

S<lt<n4 lima mlniilea 


of «!** solubilited in the recovery process. Approxijwtely 107 sW 
of HCOj” would be svailsble in the recycle strean. In rKe coagula- 
tion tank at pH 11 , this would be converted to CO 3 end precipitate 
approxlnately all of the caloiun as CoCO^. ' ApproKinncely 16.9 bH of 
C 0 CO 3 or 1690 eg/1 would be added to the solids passed onto the 

additional allowances were eado in the settling area calculations. 

The CaEdOg dose in Table S -1 is the design dose determined for 

content of the sludge without the CaCOj was determined in Chapter 4. 

The volume of sludge coaling fron the thickening operation was 
determined for a color of 5000, and was 175 ml per liter of waste 
troatod. The overflow rate from the sedimentation basin would be 
1034 gpd/ft^. The per cent solids in the settled sludge including 
the CaCOj precipitate would bo 1.996. 

flttrstlon rates nmted frga 11.7 to 20.0 lbs/ft*/hr. Th« sludse 
produced after NSSC waste treatneat would have a lower GeCDj con- 
eenttatlon than the >1elhourne sludge, thereby producing a lower fil- 
tration rate. Liptek (1074) found that coerpounds precipitated by 
Ca((M)2 can be filtered at a race of 2 to d Ibs/fc^/hr on a rotary 

vary froa 20 to 30b. A design rate of d Ibs/ft^Ar and 20b solids 

A second vacuum filter would be required if recaleination was 
Ca(CH)2 end the carbonated recycled stream. These solids would 

]b volatile solids. The fuel value 

that were previously seleoced. The doily cocal solids includiaf 
the addicionel CaCO, input to the carhonation tank would he 22,351 

5-2, 230,000 gallons of carbonated recovery liquor would he pi 

'uagneelua color removal process in Table 


clilly It SlZO/ton. 

6-1 ChemAeal Coses 

'1 of iTlagnesi™ for troaling 
a NSSC vaata uilh an initial color of SOOO aas $0,743/1000 gat. 

Line was available coonerclally ac $60/ton as of June 1976. The 

color of 5000 vaa $0,260/1000 gal. 

The cost of the polyner B35A fron Aaoricaa Cyanamid vaa $1. 50/lb 
as of June 1976. The polymer doae was 3 ng/1 and added $0,037/1000 

Lo chenical recovery. 

unless otherwise specified. The area required for settling treated 
settling basin was $18/ft^. The total capital cose for a I, 5 and 10 


•te flow would be $1S,360, $91,800 and $183,600 respectively, 
le operatlog cose of vaeiiuo filerstion without heat treatoient 

'cooval process. This cost uos $0,058/1000 ^al on a unit flow 

of etiulpoenc capable of Incinerating t> 
and 10 ogd flow was ostiaated to be $al 

if 00160181 processed increased. Approxinately 3.S, 
was cstiaatod CO be $8.52, $3.80 nod 83.50. The 

hiln. The CaCO, would not dissolve ii 

le carbonated slurry an 

Id cost t2-50/ttm or JO. 018/1000 e&I on s 

ir cJio modnesium rocoverr p>iose mi 

ted bp the author to bo J2.50D, J7,500 and 110,000 for Naste flows 

in 8 miscellaneous estiDoto that-vill be discussed later. 

The oost for recalcination at a 1 ngd plant was estiAated at 
SO. 07/1000 gal for operating cost and $700,000 capital cost. A 5 mgd 
recaloloatlon plant was eecimaced to cost $500,000 for capital enpen- 
ditures and JO.OS/1000 gal for operating cost. The ostloate for a 
10 pgd recalcination plant was $600,000 for capital expenditures and 
$0,055/1000 gal for operating cost. 

operating costs for all other equipment necessary to install the 
color renoval process. For a 1, 5 and 10 mgd effluent, the miscella* 

$0,030/1000 gal, $100,000 and $0,020/1000 gal. and $150,000 and 
$0,015/1000 gal. 

In this section of Chapter 6 the appropriete cost of.different 

ival schoBS using both Ca(OH]. and Bagncslua rocovery in 

systefis considerad 

Tho unit cost per 

dagrea of cbaaical recovery. The costs of oil 
'e presonced in Table If CalOH). recovery 

kiln Mould be eliminated. If magnesium recovery 
Ld be increesed. Coapleto chemical 

interest compounded ai 
period into the capital cosi 


■gd ms S0.44S/IOOO gll 

gal Taspectivel/. 

effluents range Tjetmon S and 10 ngd with colors varying frosi 2500 

^roaiBaCely 20iOQO gallons of NSSC waste are discharged for 

ton of N58C product produood. 
product if Che ■fl^esiuo color 




A process for resiovins io excess of 90S of the leiclal color 
TLe UBs-magnesiun color reinovBl process DfilnlV Involves color 

The untreated NSSC oaste wns shovn to have a sifnificsnc acid 
streafth. The ajsoont of Bagncsiviit and Ca(0H>2 dosed to the NSSC 

shift the optiKLCD coagulant pH. Nhen CaC0Hl2 vss used for pH control 
responslhle For color reBoval in the line-nagnesiuo color renoval 
insoluble orsano-Betallic precipiute that resioued 90L of the initial 

color. 1liis deaonscrQCcd b/ the Increase in residual color for 
■Bgnesiua additions of lO-SO na/1 to Che uaste solution. The color 

or NnOK. A precipitate was eventually produced that reaoved organic 

in the initial color of the ensce was acconpanied by a 34h reduction 

of the acid strength of the untreated hSSC sasto occurred during Che 

suggested chat Che weak ecida were responsible for most of the color 

cheBicol reaction that involved color bodies, hydroaide and magneslua 
ions which resulted in the foraacion of an insoluble precipitate. The 
wipirioal foraula of Che precipitate was HgCCIH). where R represen- 
ted the organic color bodies. 

clpltatlon in a pH dependent or a pK independent chemical reaction. 

Cationic polymers did i 

responsible for color removal in 

Ionic colored floe was greatly asals- 
; polymer. The settling of tho 

by polymer-floc bridging when 
lot by electrostatic ropulsion. 

The color reoovol process was reversibla. Reooval of tha color 
bodies froo Che sludge by Inctneracioa was required in order to reuse 
the aagnesiuB for color removal. The optimum teoperature of incinera- 
tion was 5S0®C. Increasing Che incineration tomperature from SStJ®C 

from the sludge by incineration and to dissolve the incinerated mag- 
nesium by caxbonacion. Approximately 95i of the original magnesium 
was rocovarod following three uses of the same magnesium in the lime- 

magnesium wes recycled with Che solids remaining after carbonacion. 


initial color of 

. high capital expense of 

Purthor research needs to be conducted on possible polymers tl 
can serve as settling aids. Activated silica and starch are two 
anionic polymers that are mere economical and may function as wall 
tho partially hydrolyted polyacrylamides- The NS8C waste^ should b( 

fractionated b] 

id after magnealua coagulation 

>y gel filtration b< 

be detemioed for the N55C waste. Different pulp wastes could be 

procoss can be adaptable to other pulp wastes. 

Possible areas of invostigation for additional research would be 
the structure of the Hg{0H).{s} colloids and the rate of foraacion of 
Hg(OH)*. OptltsuB rates of energy input into *the rapid mix process 

Bents. 1110 fuel value of the sludge produced in the coogulation pro- 
cess should be datoTBined and necessary experinents conducted to 
design sn incinerator for the color removal process. The solids spe- 
cies controlling siagnesiinn solubility in the carbonation process 

color removal process oa a laboratory scale aro good reasons to imple- 


ss'iiii.:;:':;:; S;; jS’S's's “ SS“ 




ss:”'ifc sr!«! •' ”"' *“■"■ "■ 
SiSiS'i:';. iss “■ “' 

^=t™i£jaS£T.s » 

ass ssissi Ss.TTJar “ ■■’" “ “ 
s- ssr ■" ■”“■■ *’ •■ ■— 

srs.:,u:'s.’"s£'?i: ■' “■’•■■“ 

rs; »sr;s 

ijuUtUl. ConsidomtSm*." JAWW. 65: 6: 417 

iss;^— £~ 


:“"i.:-i£-S'.;i'i..s~' •' ““'•' "*■■“■ ■■ ■'■ 



ss,r;..:yr ”-:;;;.;Tsr .;; si'ss.'"' “ 

S|S;. »i:'sS: .'«" EtS? " 

EtsuE-iSSiE.";- 5egsr.-;^ig- .'sr" 

asa?i«» '=£■£": snrar 

S IS'.'-Ia.X'-SS "■ '>■ 

?,s>!ka“S ;ns “• ■*" 



:; ss.s.2-a£'"' 

2r;::iS:‘,£:2',iri:: “■• “•■ •' '“■• 

Jones Shorson Taylor vos born Au|iist 24, 1941, in Hlaai, 

in Hiani, Oktahma and attended Oklahono State University in 
Stillwater, Oklahnaa on a football scholarship. Ho received the 
desree of Bachelor of Science in Industrial Enginocrins and Hanage- 
Bent in August 1965, frtm Ofclahoiaa State University. 

He accepted a position as procesa and industrial engineer with 
a Conpany in Hastings, Hinnesota fro» September 1966 until Hay 

1967. He then worked as a reaearch engineer on the Saturn program 
in Cape Canavoral, Florida for one year. In June 1966, he accepted 

Melbourne, Florida, He left Radiation in January 1971 to pursue a 
Masters degree in engineering at the University of Florida in Caines* 
vine, Florida- In Juno 1972 he received a Master of Engineering 
degree in Envlronaontal Engineering from the University of Florida. 

Ho entered the Ph.D. program in the Deportment of Environmental 
Engineering Sciences at the University of Florida in June 1972. He 
became a registered engineer in the state of Florida In April 1974. 

He is married to Janet Louisa Taylor, formerly of Melbourne, 

Florida, and has two children, James Shermsn Taylor 
Briton Ashley Taylor, ago S. Tha author presently 
Professor of Enuiranmentsl Engineering Sciences at 
Institute of Technology in Melbourne, Florida, 

I certify thet I tave reed t 
conforms to scceptable standards 
fully edc^uato, :* * 

i study end that in ay opinion 11 

a dissertation for the degree 

sccopted 05 porciol fulfillmonc of the re<:[uireiiieriC5 for the 


College of Engineering